A survey of more than 2,400 Americans published last month by HUMAN Security reveals a sharp divide in how people think about artificial intelligence in travel planning: widespread comfort with AI as a research tool, and persistent reluctance to let it act independently at the moment of payment.

Half of Americans are now comfortable with AI planning their trips

HUMAN Security released the findings on May 13, 2026, drawing on responses from adults across 48 U.S. states. The survey ran over two weeks in February 2026. Alaska and Wyoming were excluded due to insufficient responses.

The headline figure is notable. According to HUMAN Security, 54% of Americans say they are comfortable using AI to plan a vacation from start to finish. That is a majority, and it did not appear overnight. The same survey found that 47% of respondents say they are more comfortable using AI for travel planning today than they were a year ago. That shift in sentiment - from experimental to practical - tracks with wider trends that have reshaped how automated systems interact with the open web.

HUMAN Security's own infrastructure data provides a complementary angle. According to the company's 2026 State of AI Report, released on April 9, 2026, AI-driven traffic grew 187% over the course of 2025. More than 95% of that growth was concentrated in three industries: retail and e-commerce, streaming and media, and travel and hospitality. The survey findings from HUMAN Security documented on PPC Land show that automation is growing eight times faster than human traffic overall.

Travel, in particular, stands out as an environment where AI assistance delivers measurable time savings on tasks that are repetitive and research-heavy by nature.

What people trust AI to do - and what they keep for themselves

The distribution of trust across specific planning tasks is revealing. According to HUMAN Security, 64% of Americans say they would trust AI to help find local activities, making it the highest-ranked use case. Comparing prices follows at 59%, then finding food and destination recommendations at 58%. Finding lodging options sits at 51%, creating an itinerary at 48%, tracking price drops at 44%, and finding flight options at 40%.

That ranking tells a consistent story. The tasks with the highest trust scores are those where AI functions as a research assistant - sorting through options, surfacing ideas, reducing the time spent on repetitive comparison work. The tasks at the lower end of the list - flight options, price tracking - still require real money to move, even if the search work is delegated.

The time-saving argument is a significant driver of adoption. According to HUMAN Security, 49% of Americans believe AI could save them one to three hours when planning a week-long trip. A further 27% think it could shorten planning by four to six hours. With the FIFA World Cup running across 11 U.S. cities between June 12 and July 19, 2026, the survey notes that crowded destinations and rapidly shifting prices make these efficiency gains feel more tangible during peak travel periods.

When respondents were asked to identify the biggest benefits of using AI for summer travel planning, saving time came first at 34%, followed by making planning easier at 32%, discovering new destinations at 28%, and finding better deals at 26%.

The approval checkpoint divides comfortable from uncomfortable

The clearest fault line in the survey data appears at the moment of booking. According to HUMAN Security, 43% of Americans would be comfortable letting an AI assistant book travel on their behalf if the system asked for final approval before completing the payment. That is nearly four in ten people prepared to hand over some booking authority, as long as a human checkpoint remains in place.

Remove that checkpoint, and the picture changes entirely. Only 12% of respondents say they would be comfortable allowing an AI assistant to book travel deals automatically, without requesting approval each time, even when the AI operates within a budget and set of preferences the traveler has established. The share who would be uncomfortable with that level of autonomy is 82%.

That gap - between 43% comfortable with supervised booking and 12% comfortable with fully autonomous booking - is one of the sharpest contrasts in the survey. It suggests the approval step is not a minor procedural detail. For most Americans, it is the boundary that separates an AI assistant from an AI actor.

The payment dimension adds another layer. According to HUMAN Security, only 18% of Americans would trust AI systems to securely handle their payment information. 58% would not. Planning support may feel useful and low-risk; payment feels different.

HUMAN Security's April 2026 update on agentic traffic patterns found that most observed AI activity still centers on product and search routes, while activity in account, authentication, and checkout environments continues to emerge but remains at an earlier stage of development. Consumer sentiment in this survey aligns with that trajectory: people are more comfortable with AI doing the browsing than with AI completing the transaction.

Human agents still outrank AI for important trips

The survey asked respondents to consider not just any trip, but an important one. On that question, 82% say they would trust a human travel agent more than an AI travel assistant to plan and book. Only 18% would trust the AI more for a high-stakes booking.

There is one exception worth noting. Among people who travel very often, 31% say they would trust the AI more. Frequent travelers also show higher adoption rates overall: 68% of those who travel very often have used AI for vacation planning, compared to 21% of those who travel rarely. Repeated exposure to the friction points of travel planning - price comparison, booking windows, itinerary building - appears to lower the threshold for accepting AI assistance.

That same group also shows greater willingness to allow autonomous action. According to HUMAN Security, 34% of very frequent travelers say they would press a button allowing AI to book on their behalf, and 13% say they would do so without hesitation.

Who carries the blame when things go wrong

The accountability question sits at the center of how Americans think about AI booking errors. According to HUMAN Security, 40% of respondents say the company that built the AI tool should be held responsible if an error costs money or disrupts a trip. 20% say the user should bear responsibility. 19% say responsibility should be shared.

Among those who favor shared responsibility, the breakdown is instructive. 83% say the AI company should share in the blame. 72% point to the user. 55% include the platform where the booking occurred. 31% extend responsibility to the travel provider itself.

Americans are not treating AI booking errors as isolated technical glitches. The accountability chain they describe is layered - involving the tool's creator, the platform, the user, and the underlying travel provider. That expectation of distributed accountability is worth watching as autonomous booking systems develop, because it suggests consumers will hold a broader set of actors responsible than existing terms of service may anticipate.

The accountability dynamics here connect to infrastructure changes happening across the agentic commerce ecosystem. Experian, Visa, and Cloudflare have each launched or expanded identity verification frameworks specifically designed to create traceable links between consumers and the AI agents acting on their behalf, addressing exactly the kind of responsibility gap the HUMAN Security survey data highlights.

Personalization remains a contested promise

One of the more unexpected findings concerns how people view AI's ability to personalize a trip rather than flatten it. According to HUMAN Security, 43% of Americans think using AI to plan vacations would make a trip more generic. That skepticism about personalization is higher than might be expected for a technology often marketed on its ability to tailor recommendations.

The data on stored preferences reinforces a similar divide. According to the survey, 43% say they would be comfortable with an AI assistant storing and using past trip history, preferences, and spending patterns to tailor future bookings. 39% would be uncomfortable with it. The near-even split suggests that personalization is not a settled value proposition in travel. Many people can see its potential usefulness. Almost as many are uneasy about what the required data retention involves.

This tension - between the convenience of personalized AI and discomfort with storing sensitive behavioral data - has been documented in broader consumer trust research. A July 2025 study by Usercentrics found that 59% of consumers express discomfort with their data being used to train AI systems, and the hospitality and travel industry ranked among the sectors with lower consumer trust in data handling, with only 22% of consumers expressing trust in how travel companies manage their information.

Regional variation adds texture to national averages

The survey also collected state-level data. Nationally, 42.6% of respondents said they would be comfortable allowing AI to book travel with final approval before payment. Several states came in well above that national average. According to HUMAN Security, Alabama reached 64.0%, Michigan 59.2%, Missouri 58.0%, New Jersey 54.0%, and Connecticut 49.1%. Alabama's figure sits more than 21 percentage points above the national mark.

The picture changes when final approval is removed. Nationally, only 11.7% said they would accept fully autonomous booking. Even in the states most open to that level of AI authority, the numbers remain low. According to the survey, Texas came in at 20.8%, Georgia at 18.8%, Michigan and Virginia each at 18.4%, and Florida at 18.0%. Texas leads this category, but even there, one in five people represents the ceiling of comfort with autonomous action.

Michigan appears near the top of both lists - more comfortable than average with both supervised and autonomous booking - suggesting that within individual states, the factors driving AI adoption in travel may reinforce each other across categories.

For state-level task preferences, the survey found that finding local activities or attractions was the top-ranked AI use case in 25 states. Comparing prices led in 10 states. Recommending destinations was the top choice in 8 states, and finding restaurant recommendations in 5 states. In half the country, the most trusted AI use case for travel is figuring out what to do once people arrive at their destination - not deciding whether to go.

What the numbers mean for the marketing community

For advertising professionals tracking AI's trajectory across digital commerce, the HUMAN Security survey adds a behavioral dimension to the infrastructure data. AI-driven traffic in travel and hospitality is rising fast. But consumer attitudes toward the tools generating that traffic are layered and conditional.

The agentic commerce pattern documented across multiple sectors in 2025 and 2026 has shown that consumer readiness for autonomous AI action lags the technology's technical capability. Lipsman's probability analysis of agentic shopping - which estimated compounding low probabilities across multiple necessary behavioral conditions - finds a practical parallel in travel. Even when consumers accept AI for research, the majority require an approval step before money moves.

That gap matters for anyone building, deploying, or advertising through AI-mediated travel channels. Google's integration of AI travel tools into Search campaigns represents one layer of this shift. The expansion of AI booking features across 200 countries in November 2025 represented another. Both moves assume growing consumer comfort with AI-mediated travel, which the HUMAN Security data suggests is real but not yet unreserved.

The accountability framework consumers expect - AI companies, platforms, users, and travel providers all sharing responsibility - will eventually require clearer rules than currently exist. Research showing that one-third of brands risk damaging trust through premature AI deployment points to the same pressure: the speed of AI adoption in travel infrastructure does not automatically produce consumer confidence in the results.

The HUMAN Security survey's February 2026 fieldwork precedes the FIFA World Cup travel surge. Whether the event - and the planning pressure that comes with it - accelerates adoption or surfaces new friction points for AI-assisted booking systems will be a practical test of the consumer attitudes the data describes.

Timeline

Summary

Who: HUMAN Security, a cybersecurity company that processes more than 20 trillion digital interactions weekly and operates the Human Defense Platform. The survey was conducted with a sample of more than 2,400 Americans across 48 U.S. states.

What: A survey measuring U.S. consumer attitudes toward AI in travel planning and booking, covering comfort levels with AI-assisted itinerary building, autonomous booking, payment handling, personalization, data storage, and accountability. Key findings include 54% comfort with AI full-trip planning, 43% comfort with AI booking given final approval, 12% comfort with autonomous booking without approval, and 18% willingness to trust AI with payment information.

When: The survey fieldwork ran over two weeks in February 2026. HUMAN Security published the results on May 13, 2026.

Where: The survey covered adults across 48 U.S. states (Alaska and Wyoming were excluded). Regional variation was analyzed at the state level, with Alabama, Michigan, Missouri, New Jersey, and Connecticut showing above-average comfort with supervised AI booking, and Texas, Georgia, Michigan, Virginia, and Florida leading on autonomous booking comfort.

Why: HUMAN Security conducted the research to document how consumer attitudes toward AI-driven travel planning are evolving as agentic AI traffic in travel and hospitality grows rapidly. AI-driven traffic grew 187% in 2025 according to HUMAN Security's own platform data. The findings are intended to inform how organizations understand the gap between AI's technical capabilities in travel and the trust conditions under which consumers are actually willing to use them.